The effects of HIP treatment on the thermoelectric properties, including the Seebeck coefficient, electrical conductivity, and thermoelectric power factor, of the samples were examined. Figure 1 shows the temperature dependence of the thermoelectric properties of the samples following HIP treatment. The absolute value of the Seebeck coefficient is shown in Fig. 1(a). Although the Seebeck coefficient was reduced by approximately 50% before and after HIP treatment, its value remained independent of the HIP treatment temperature and exhibited nearly constant behavior, as shown in Fig. 1(a). In contrast, the electrical conductivity substantially increased with increasing HIP treatment temperature, as shown in Fig. 1(b). The power factor, which is the ratio of the Seebeck coefficient to the electrical conductivity, increased with increasing temperature during HIP, as shown in Fig. 1(c). The improvement in the thermoelectric properties was attributed to the HIP treatment, despite the decrease in the Seebeck coefficient. In a previous study, it was reported that the generation of oxygen vacancies occurred in a sample following HIP treatment, resulting in a change in color from white to light gray.17 In this study, the significant increase in conductivity suggested that oxygen vacancies were generated within the samples. To confirm the presence of oxygen vacancies in the HIP-treated samples, the phase identification of the samples before and after HIP treatment was analyzed via X-ray diffraction at room temperature. However, the X-ray diffraction results indicated that the sample after HIP treatment exhibited the same crystal structure as the sample before HIP treatment. Therefore, the oxygen vacancies introduced by HIP treatment may not be detectable by X-ray diffraction. We investigated the effects of the processing atmosphere, processing pressure, and heaters during HIP treatment to identify the factors contributing to the oxygen vacancies due to HIP treatment.
The influence of gaseous species in the HIP treatment was verified. Figure 2 shows the thermoelectric power factor of the ZnO samples after HIP treatment with a mixture of argon and oxygen or argon gas. As shown in Fig. 2, there are variations in the Seebeck coefficient, electrical conductivity and power factor of the samples after HIP treatment under both gas conditions. In particular, the electrical conductivity and the thermoelectric power factor of the sample after HIP treatment under argon gas improved. To clarify this factor, the relationship between the effect of HIP treatment and the density of ZnO was investigated. The density of the samples was measured via Archimedes’ method. Figures 3(a), 3(b) and 3(c) show the volume, mass and density of the samples before and after HIP treatment at 1273 K under argon gas or under a mixed gas of argon and oxygen for 2 h, respectively. As shown in Fig. 3, the sample after HIP treatment under argon gas was denser than the sample after HIP treatment under mixed gas. This occurred because the degree of reduction in the volume of the sample was large regardless of the decrease in mass during the HIP treatment under argon gas. Therefore, the atmosphere of argon gas was considered one of the factors leading to an increase in the number of oxygen vacancies, confirming that HIP treatment was effective in densifying ZnO.
We compared the thermoelectric power factor between HIP treatment under high pressure and heat treatment under atmospheric pressure to verify the effect of pressure under argon gas. Figure 4 shows the thermoelectric power factor of the ZnO samples after HIP treatment or heat treatment. As shown in Fig. 4, the electrical conductivity and thermoelectric power factor improved after both HIP treatment and heat treatment, indicating that HIP treatment is more effective than heat treatment. Therefore, the high treatment pressure was one of the factors leading to an increase in the number of oxygen vacancies.
Since the reduction effect of graphite heaters in HIP equipment has been reported, we investigated the effect of the heater material to determine the generation factor of oxygen vacancies. Figure 5 shows the thermoelectric properties of the samples before and after HIP treatment with a graphite heater or a molybdenum (Mo) heater. However, the results when the graphite heater was used were almost the same as those when the Mo heater was used, as shown in Fig. 5. This finding indicates that the generation of oxygen vacancies is independent of the type of heater material.
We performed temperature annealing tests in an oxygen atmosphere to evaluate the thermal reliability of the HIP-treated ZnO samples. Oxygen annealing temperatures were carried out in the range of 773 K to 1273 K at atmospheric pressure. Figure 6 shows the temperature dependence of the effect of oxygen annealing on the thermoelectric properties of the HIP-treated ZnO samples. As shown in Fig. 6, no change in the thermoelectric power factor was observed up to 873 K, but above this temperature, the thermoelectric power factor decreased with increasing temperature due to the decrease in electrical conductivity. In other words, it was confirmed that the HIP-treated ZnO samples can stably maintain a current value of up to 873 K in the atmosphere.